Geek Culture
Published in

Geek Culture

Improve Neural Networks by using Complex Numbers

Can Complex Functions be the next breakthrough in Computer Vision?

Understanding Convolutional Neural Networks

The feature extraction is the true CNN revolution. Taken from IBM’s Writeup on ConvNets
CNNs use the sliding window technique to build their feature maps. As you can see, Good Machine Learning requires good software engineering. Image Source

Enter Hybrid Neural Networks

In a hybrid neural network, the expensive convolutional layers are replaced by a non-trainable fixed transform with a great reduction in parameters.

The basic idea behind Hybrid NNs and this paper

The amazing CoSh Network

If we can understand what makes these amazing results tick, we can create much better solutions.

The Magical Properties of Complex Functions

Image Source
If you have experience with split-ReLU, let me know.

Phase Congruency

Gradients fluctuate wildly across scale but phase remains very stable at critical parts of the image. This makes phase a great base for detecting important features.
“Fig 4 shows despite the considerable perturbations (blurring and Gaussian noise), CoShRem remain stable to most of the characteristic edges and ridges (two step discontinuity in close proximity).”

Reach out to me



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Devansh- Machine Learning Made Simple

Deep Insights about Artificial Intelligence (AI), Machine Learning, Software Engineering, and the Tech Industry. Follow me to come out on top